By mid-March, the backend prototype and command-line pipeline were functioning end to end. That changed the nature of the work. The technical core existed. Interface wiring, pilot usability, repeatable controls, and deployment hardware became the real blockers.
The backend crossed the first proof line
The local pipeline could ingest material, retrieve context, send that context through a local reasoning path, and return work through the command line. That proved the system was more than a presentation concept.
It carried the label working MVP on purpose. A client cannot live in a founder's terminal, infer the right document-handling rules, or accept an undefined maintenance process. The proof shifted the bottleneck from core feasibility to productization.
UI wiring became a business problem
Pilot users needed a safe way to choose a collection, ask a question, inspect sources, understand uncertainty, and know what happened to their data. Each missing interface state raised the training burden and made failure harder to interpret. The unfinished interface was a business problem, not a cosmetic one.
The decision was to focus on pilot usability instead of adding model features for their own sake. A narrow workflow a real professional could evaluate on the merits beat a long feature list that only worked when the founder operated it.
Separate demonstration hardware from client deployment
Ralph's own rig handled controlled demonstrations and early discovery well enough. Leaving a personal development machine inside a client environment was a different matter. That distinction forced an early conversation about dedicated hardware, asset ownership, support, backup, physical control, and what happens at the end of a pilot.
The initial target became a small group of three to five firms rather than a broad launch. A focused founding cohort could expose workflow failures and security assumptions before the company tried to standardize a larger deployment package.
Fund validation before scale
The near-term funding target was roughly $45,000 to $50,000 for pilot hardware and deployment, well short of a venture round. MSU Research Foundation, the Technology Innovation Center and SmartZone ecosystem, LEAP, and the Michigan SBDC commercialization path became part of the operating strategy.
The company chose customer discovery, warm introductions, pilots, and Michigan ecosystem support over premature venture pressure. The money was meant to buy proof: dedicated systems, repeatable deployment, and evidence that a private workflow created enough value to justify the next step.